This seems like a scenario where perfect is getting in the way of good enough. If the measure uses "medication, ordered" then you run the risk of including patients who didn't actually have the medication dispensed. However, if the measure uses "medication, dispensed" then you run a much higher risk of not including patients who did actually have the medication dispensed, just because you don't have the data. Either way the measure is going to have some level of inaccuracy - which is more egregious?

Floyd's information above seems to suggest that the dispensing data will become more available as the standards mature and are more widely adopted. Perhaps the short term preference should be for "medication, ordered" and once the data becomes more widely available it could be changed back to "medication, dispensed"?

We also support the change to move away from Medication, Dispensed, due to the availability and logistics of adding this information into a patient's chart for it to be available for eCQM reporting.

Dispensing data is clearly closer to understanding if a patient has received the medication (although not that the patient took the medication). The HIT Standards Committee looked at this issue and I asked from input from one of the chairs of the Interoperability Standards Advisory Task Force of the HITSC. The following response suggests work needs to be done to fully understand the workflow and to encourage adoption of existing standards in a 'standard' way before dispensing information will be available:
The NCPDP Script Implementation Guide includes 2 transactions of interest: (a) Prescription fill status (RxFill) which is not widely adopted and (b) Mediation History which requires a consent process.

The Interoperability Standards Advisory Task Force of the Health IT Standards Committee advised:
• [R]NCPDP SCRIPT Standard, Implementation Guide, Version 10.6 is the best available standard for creating and transmitting a new prescription in the outpatient setting.
• We would advise caution in including all message transaction within the NCPDP SCRIPT Standard as workflows and system capabilities have not been vetted well in real practice.
• There are two message transactions in NCPDP SCRIPT v10.5 that we are in agreement with considering: Cancel Prescription (CANRX, CANRES) and Refill Prescription (REFREQ, REFRES), which could better facilitate prescriber-pharmacist communications.
The Content Standards Workgroup also provided the following advice to the HITSC:
• In order to improve the ability to capture and represent medication history consider prioritizing the additional transactions or segments, namely Change Prescription, Refill Prescription, Cancel Prescription, Fill Status, and Medication to address needed functionalities in both the pharmacy software systems and the prescriber’s software systems / health IT modules.
• Agreement with adding Cancel Prescription (CANRX, CANRES) and Refill Prescription (REFREQ, REFRES) as transactions or segments from NCPDP SCRIPT v10.6 in order to better facilitate prescriber-pharmacist communication
• RXCHG: we are aware that several pharmacy software vendors have this functionality to be developed by the end of 2015 and in 2016, so this would be a criterion that could be rapidly adopted if both prescribers and dispensers have the functionality.
• Fill Status (RXFILL) notifications should occur in the electronic health record and not outside (e.g., email or other means of communication). However, we feel we need healthcare professional input (e.g., from those engaged in prescribing and dispensing prescriptions) on what the appropriate triggers should be for notifications to ensure criterion does not generate excessive messaging. This specifically applies to Fill Status (RXFILL).

We support the change to make the measure logic check for “medication, ordered” as we find the dispense data to be very sparse.

This issue is under discussion in the Change Review Board.

I think no one doubts that if you asked providers and provider organizations, a large majority would want this measure to look at orders. This is because dispense data is very frequently unavailable in EHR implementations in 2015, and because it relies on actions that are unlikely to be under the direct control of the provider.

Without access to this data, a provider's performance rate on the measure is equal to what it would be if the provider gave poor clinical care. The ramifications of this should not be lost on anyone.

The Change Review Process (CRP) discussed the issue of “medication dispensed vs. ordered” that affects the following measures:
• Use of Appropriate Medications for Asthma (CMS126)
• Follow-Up Care for Children Prescribed Attention Deficit/Hyperactivity Disorder (ADHD) Medication (CMS136)
• Anti-depressant Medication Management (CMS128)

The CRP discussed the options below and we would appreciate additional stakeholder feedback on them. During the November CRP meeting, NCQA will present a summary of the public comments, make our recommendation regarding the issue, and the CRP will vote on NCQA’s recommendation.
• Measures stay as “dispensed”
• Measures are changed to “ordered”
• Measures include stratification, where logic would look for “dispensed” first, and if not available, then look for “ordered”

We welcome all public comment for the next two weeks prior to the November CRP.

Thank you!

This seems to be essentially a duplicate of CQM-795